ECONOMIC ANALYSIS Budget Magic and the Social Security Tax Cap By Martin A. Sullivan martysullivan@comcast.net Raising the annually adjusted cap on wages subject to the Social Security tax in 2005 from $90,000 to $150,000 would increase federal revenue by about $50 billion in 2010; over the long run, it would cut the Social Security Trust Fund s fiscal shortfall in half. Eliminating the cap could raise more than $100 billion by 2010 and make Social Security s financial problems disappear for the next 75 years. Those are important facts to keep in mind now that frustrated White House officials have hinted that President Bush may retreat on his no-tax-increase pledge for fixing Social Security. But there s one other important thing to know: Any FICA tax increase would be a two-for-one budgeting special. That s because politicians allow official scorekeepers to double count. Social Security funds are commingled with other government funds in the calculation of the total deficit. (Government economists also calculate a far better measure of federal finances the on-budget deficit, which excludes Social Security receipts but nobody pays any attention to that.) Because the Social Security system is currently running a surplus, using the total deficit when talking about the deficit makes the fiscal sins of the president and the Congress appear less troubling and that s troubling in itself, considering how bad the current figures look. Although that shady accounting practice makes overall deficits look smaller, it usually has little immediate effect on deliberations over specific budget proposals. Raising the cap on Social Security payroll taxes is an exception to that last rule. Here s where the twofer comes in. Under the pretzel logic of the current system, a $50 billion increase in Social Security revenue does two things: (1) it improves the Social Security Trust Fund finances by increasing the fund s income by $50 billion, and (2) in calculating the total deficit of the federal government, it makes the deficit look $50 billion smaller which provides a $50 billion cushion for spending increases and tax cuts. In other words, Congress could raise the cap on income subject to Social Security taxes (thereby increasing Social Security taxes to help save Social Security) and at the same time cut income taxes by an equal amount, and there would be no change in the overall deficit. Voila! It is a sign of the times that there are highly intelligent people on Capitol Hill who are wondering if they can really get away with that. The rest of this article talks about other policy issues involving an increase in the Social Security payroll tax cap. But the importance of the policy issues discussed here and the political issues could all become secondary factors in the FICA cap debate if a deal can be made that includes the two-for-one special. Behind the Numbers The Social Security Administration (SSA) estimates that in 2005, 84.7 percent of all wages will be below the taxable maximum of $90,000. That percentage was 90 percent in 1983, but it has declined substantially since then because wage growth has been greater for high-wage workers than for low-wage workers. The disparity in wage growth is expected to continue, and the percentage of wages taxable is expected to drop to about 83 percent in 2013 and remain there in the following years. This is all illustrated in Figure 1 on p. 1248. By raising the cap, Congress could save Social Security without any net tax increase, benefit cut, or increase in the deficit. Figure 2 on p. 1249 shows how much money is on the line. The SSA estimates that in 2005 there will be $860 billion in wages paid in excess of the $90,000 wage cap. The Social Security tax rate is 12.4 percent, split equally between employer and employee. If there were no cap on taxable wages, the potential revenue increase for 2005 would be $107 billion. We say potential because (as explained below) there are many reasons for expecting that the amount would not be fully realized. By 2015 taxable wages above the cap are expected to increase to $1.5 trillion. The corresponding amount of potential revenue is $187 billion. Let s put money aside and talk about the people who would pay the additional taxes. The latest year for which detailed data are published is 2001. That s a little old, but it gives a pretty good picture of those who would be affected. In that year, the wage cap was $80,400. Of 153.8 million individuals paying Social Security tax, 9.1 million had maximum taxable wages. Approximately 80 percent of those were men. Most of those workers received only wage and salary income. Fewer than 10 percent received any self-employment income. On average, each of the 9.1 million individuals had an income of $160,300. If the cap on taxable wages of $80,400 had been eliminated in 2001, the average Social Security tax increase on each of those individuals would have been slightly less than $10,000. Averages don t mean much in this case because the distribution of earnings above the cap is highly 1246 TAX NOTES, March 14, 2005
Doc 2005-4972 (5 pgs) Estimated Effect on Revenue and Trust Fund Actuarial Balance of Raising the Cap on Wages Subject to Social Security Tax Current Law SSA Estimates JCT Estimates % of Total Wages Taxable 84 90 87 100 90 91 92 Taxable Wage Cap (in thousands of dollars) 90 147? None 150 170 190 Estimated Revenue Increase in 2010 (in billions of dollars) 55.9? 145.2 45.2 53.7 60.5 Social Security Trust Fund Actuarial Balance (as % of current taxable payroll) -1.91-0.85-1.48 >0??? % Reduction in Trust Fund Shortfall 55 22 >100??? Sources and notes: Provided at the end of this article. skewed. We would like to show you what that distribution was, but the information compiled from a sample of more than one million Forms W-2 is not published by either the IRS or the SSA. The numbers are available, however, to revenue estimators at the SSA and the Joint Committee on Taxation who use them to estimate the effect of raising the payroll cap (for example, to $150,000) but not eliminating it. Tricky Estimates It is easier to estimate the effect of eliminating the cap than to estimate the effect of raising it, because elimination does not require any knowledge of the distribution of earnings. Unfortunately for estimators, politicians are not talking about eliminating the cap. The specifics of the proposals under discussion are not clear, but press reports have associated the name of Sen. Lindsey Graham, R-S.C., with proposals to raise the taxable earnings cap to $150,000 or $200,000. Estimators often finesse the difficulty of relating amounts of revenue to increases in the cap. Instead of telling us what the new capped amount would be, they simply say the new capped amount would raise the percentage of taxable wages to 87 percent or 90 percent and leave it up to someone else to figure out exactly what the dollar amount of the cap would have to be to achieve that. For example, as part of its estimate for one reform plan, the SSA estimated the effect of raising the taxable maximum so that taxable wages remained at a level equal to 87 percent of total wages, but it did not indicate what that the dollar amount of that cap might be in any year. (Citations are provided in the notes at the end of the article.) There are some publicly available estimates that relate revenue increases to higher dollar levels in the cap. They are summarized in the table above. The two most important aspects of the table were noted in the first paragraph of this article: (1) raising the cap to $150,000 would cut the long-term Social Security actuarial imbalance in half, and (2) eliminating the cap would eliminate the imbalance. Now let s take a more thorough look at the estimates. You probably noticed that the SSA s estimates of the revenue effect of raising the cap are larger than the JCT s estimates. There are several reasons for the differences between the two sets of estimates. First, the SSA estimates assume that the maximum taxable amount will be adjusted each year to keep taxable wages at a constant percentage (for example, 90 percent) of total wages. The JCT estimates a proposal in which taxable revenue reaches a targeted percentage of wages (again, for example, 90 percent) in the first year but then adjusts the cap as under current law, where it is indexed to average wage growth in the economy. Therefore, under the proposal estimated by the JCT, the percentage of wages subject to taxes would decline (as it will under current law). Second, the SSA estimates look solely at trust fund revenue and, under trust fund accounting, reports its estimates on a calendar-year basis. The JCT estimates include the effect on all sources of revenue (income and Medicare taxes are likely to be affected) and, in accordance with federal government accounting, reports its estimates on a fiscalyear basis. The SSA estimates always include effects on the trust fund s actuarial balance; the JCT estimates never do. Finally, the smaller JCT estimates assume that reported wages will decline when tax rates increase. Income and Medicare Tax Offsets As all self-employed individuals know, part of the bite of FICA taxes is reduced because the employer portion of those taxes is deductible for income tax purposes. So for a self-employed individual in the 35 percent income tax bracket, an additional 12.4 percent tax effectively amounts to a 10.2 percent increase. That s because for each dollar TAX NOTES, March 14, 2005 1247
91% 90% 89% 88% 87% 86% 85% 84% 83% 82% Figure 1 Percentage of Wages Subject to Payroll Tax, 1980-2020 90.0% 85.8% 87.8% earned, income taxes are reduced by 35 percent of 6.2 cents, that is 2.2 cents. Like the taxpayer, the JCT when scoring effects on total revenue looks at the net amount of 10.2 percent. But the SSA just cares about the payroll tax, and it registers the full 12.4 percent as a tax increase. There can also be negative effects on Medicare taxes from increases in Social Security taxes. As described below, for a variety of reasons there could be a reduction in wages if the Social Security wage cap is increased. Because Medicare taxes are not subject to a cap, the behavioral responses induced by broader application of the Social Security tax would also reduce the Medicare tax base. And so increases in the Social Security tax cap would have a detrimental effect on the Medicare Trust Fund. Personal Goodwill From an economic perspective, raising the maximum amount of wages subject to the 12.4 percent Social Security tax is a particularly harmful proposal. Let s say the proposal is to raise the cap to $200,000. Although that would raise taxes for all of the 10 million or so taxpayers with wages above $90,000, economists worry mostly about those individuals earning between $90,000 and $200,000 who 85.1% 83.3% 84.7% Actual Projections 83.8% 83.3% 1980 1985 1990 1995 2000 2005 2010 2015 2020 Sources: See notes at end of article. would experience increases in their marginal tax rate. Because additional work effort would generate less after-tax income, there is concern that labor supplied to the economy would shrink. Of course, the amount of that reduction is unknown and is the subject of vigorous debate. Whatever the effect on labor supplied from raising Social Security taxes, the effect on wages reported is likely to be much larger. In the context of previous increases in the marginal tax rate, economist Martin Feldstein of Harvard and the National Bureau of Economic Research has argued that tax avoidance can have a large impact on expected revenue. Here are two ways to reduce payroll taxes. First, instead of operating as a sole proprietorship, a self-employed individual can incorporate the business and elect subchapter S status. All of the sole proprietorship s income would be subject to payroll tax. An S corporation could pay its employee-owner a reasonable salary and pay the remaining business profit in the form of dividend not subject to payroll tax. That is exactly what former Democratic vice presidential candidate John Edwards did when he made more than $20 million in his law practice. His 1248 TAX NOTES, March 14, 2005
Doc 2005-4972 (5 pgs) Billions of Dollars 1,600 1,400 1,200 1,000 800 600 400 200 0 $860 Wages Potential Revenue $919 $975 $1,041 $1,139 S corporation paid him a salary of $380,000, and the remainder escaped the 2.9 percent Medicare payroll tax. Various tax advisers on the Internet advertise another way to avoid payroll taxes. A high-wage employee claims to be the owner of personal goodwill, which is then made available to the employer for a fee. The formerly taxable wage income is converted into royalty income not subject to payroll tax. $1,173 $1,241 $1,313 $1,383 $1,447 $1,512 $107 $114 $121 $129 $141 $145 $154 $163 $172 $179 $187 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Sources: See notes at end of article. Figure 2 Taxable Wages Above the Cap and Potential Revenue From Eliminating the Cap Fuzzy Math Politicians and the press like to use numbers, but routinely they are imprecise about what the numbers mean. In this article, we have tried to sort out some of the estimating issues suggested by proposals to increase the payroll tax cap. Even if estimates are published and well documented, no one set of numbers can answer all the relevant policy questions. To assess the fiscal effect of raising Social Security taxes, you need SSA estimates for measuring the effect on the trust fund and JCT estimates for measuring the effect on the federal budget. TAX NOTES, March 14, 2005 1249
Notes and Sources The total deficit in fiscal 2004 was $412 billion. The on-budget deficit (which excludes current Social Security surplus) was $567 billion. (Congressional Budget Office, The Budget and Economic Outlook: Fiscal Years 2006 to 2015, January 2005, Table 1.) Information about the number of individuals earning wages at or above the taxable maximum and about the amount of earnings subject to tax, including historical information presented in Figure 1, is from the Social Security Administration s Annual Statistical Supplement, 2003, released July 2004, tables 4.B1 through 4.B9. Projections in Figure 1 from 2005 onward are based on data from a memo dated July 19, 2004, from SSA actuaries Chris Chaplain and Alice Wade to Stephen C. Goss, chief actuary, entitled Estimated Long-Range OASDI Financial Effects of a Proposal to Increase the Taxable Maximum Earnings Level. The estimates reported in the memo were produced at the request of the staff of former Rep. Joseph Hoeffel, D-Pa. (who contested the reelection of Sen.Arlen Specter, R-Pa., in 2004 and lost). The estimate is not available on the SSAWeb site but is on the Web site of a policy-oriented nonprofit organization (http://www.centristpolicynetwork.org/legis lative_updates/files_2004/oact_taxcap_071904. pdf). Data from the SSA on the taxable percentage of total wages for 2003 and 2004 could not be located, so the available information for 2002 and 2005 was interpolated in Figure 1. The interpolated figure for 2003 of 85 percent is lower than the figure of approximately 86 percent reported by the Congressional Budget Office in its February 2005 Budget Options (full cite given below). An estimate for increasing the percentage of taxable wages to 87 percent of all wages (referred to in the text and in the table) is provided in: Social Security Administration, Estimated OASDI Financial Effects of the Bipartisan Retirement Security Act legislation subsequently introduced as H.R. 3821 (108th Congress) by Representative Jim Kolbe and Representative Charles Stenholm, Feb. 11, 2004, available at http://www.ssa.gov/oact/solvency/ Kolbe_20040211.html. This estimate is based on 2003 assumptions and is not strictly comparable with the other SSA estimates presented in the table, which are based on 2004 assumptions. The SSA estimate in the table for increasing the percentage of taxable wages to 90 percent is from the July 19, 2004 Hoeffel memo. The table also reports an SSA estimate for total elimination of the payroll tax cap. That is really an estimate made by the author based on projections from the Hoeffel memo and is the same as shown in Figure 2. Estimates of potential revenue in this figure are based on projections from the memo of total wages and taxable wages. To arrive at the shown estimates of potential revenue in each year, the difference between these two figures is multiplied by 0.124. For the SSA, actuarial balance is the condition of the Social Security trust fund being solvent for the next 75 years. Shortfalls are expressed in percentage of taxable payroll. According to the latest (2004) estimates available (using intermediate assumptions), the actuarial imbalance is 1.89 percent of payroll. That means that an immediate and permanent increase equal to 1.89 percent of taxable payroll in every year would restore the trust fund to 75-year solvency. Given that the current tax rate is 12.4 percent, that is equivalent to a 15.3 percent increase in payroll taxes. We can say that elimination of the cap restores the trust fund s actuarial balance because that results in a tax increase varying from 18 percent to 20 percent (depending on the year). The JCT estimates reported in the table were not published by the JCT but by the Congressional Budget Office. They are part of the description of revenue option 39, Increase the Upper Limit for Earnings Subject to the Social Security Payroll Tax. See Congressional Budget Office, Budget Options, February 2005, at http://www.cbo.gov. 1250 TAX NOTES, March 14, 2005